Analysis of tool wear characteristics and estimation

21 April 2021

Tool wear characterization

The pre-worn tools were studied using SEM microscopy at different points along the cutting edge and flank wear values were measured, exemplified in Fig.1. The crater wear (KT) was measured using Alicona Infinite Focus 3-D optical microscope by comparing a new edge with a used edge and measuring the distance between surface levels (Fig.2). The results are summarized in Fig.3.


pcd and pcbn cutting tools

Fig. 1. SEM topography of a worn cutting edge (90 pieces, tool 3).

pcd and pcbn cutting tools

Fig. 2. 3-D optical microscopy of flank wear land (VB) and rake crater.

pcd and pcbn cutting tools

Fig. 3. Progression of flank wear and crater depth with a number of machined workpieces.


Tool wear estimation

A multitude of strategies is available for the development of estimative/predictive models. The approach presented in this study is based on the process setup of one industrial representative but may be applied to machining cells and/or lines of other automotive companies, those who have issues with the underutilization of cutting tools due to batch production and variation of process parameters. A flowchart illustrating the approach for the optimization of tool utilization is presented in Fig.4.


pcd and pcbn cutting tools

Fig. 4. Flowchart for prediction of tool wear and optimization of tool utilization.

The lower branch of the flowchart (blue region) represents a typical strategy for the estimation of the remaining tool life. The blocks in this branch might be performed automatically or manually based on the observations of tool condition and product quality and accordingly a decision about the tool change. This approach works well for a steady process when the tool life is almost constant, and this stresses the accuracy of anomaly detection block as the most important.

Such a TCM system allows detecting the tool breakage/damage (Anomaly detection  and predicting a tool life through the signal processing and usage of some models. This approach is quite simple and efficient providing that the process is “stable” with minimal variations in the process parameters. The presence of variations (for instance, in the workpiece or/and tool material properties) affects the accuracy of the tool life assessment resulting in the decrease of tool utilization time.

The upper branch in Fig.4 (green region) has a supportive function to the basic one (it works simultaneously with or embedded into the TCM system) aiming to increase the accuracy of the estimation of remaining tool life and improve the process controllability and can be built through a deeper understanding of the process's specifics. For this purpose, the process is divided into three parts, and signals at the beginning (Signal 0) and end (Signal 2) of the process are involved in additional computations and analysis. Preprocessing of the signal(s) from the sensors when the operation began allows the estimation of the workpiece condition (e.g., hardness, composition, etc.) using the data on the state of the tool from the previous workpiece machined. These data are used further for the prediction of tool life (estimation of remaining tool life) using certain empirical model (e.g., Tailor, Colding, ML-based, or recommended by tool manufacturer).

Signal 2 at the end of the operation is used for the estimation of the current state of the cutting edge. The comparison of the predicted and estimated state of the tool allows for correction of parameters of the used model those might vary with tool wear development. Also, it may indicate the change of wear mechanism due to, for instance, the variation of workpiece material composition. The data obtained allow an estimation of the remaining tool life and making decision regarding further use of the tool. The information on the state of the process is updated at the end of the cycle.

Following the steps of the presented flowchart, the next problems should be solved: estimation of workpiece condition (hardness in our case), tool wear prediction with use of data on hardness, tool wear estimation based on the signal only, all of which are within the scope of current study. 


Conclusion can be drawn from the obtained results

Tool wear is characterized by two parameters (flank wear land VB and crater depth KT) that are typical for hard turning operations with PCBN tools. Crater formation is the dominating wear mechanism at the beginning of the process (up to 60 out of 120 workpieces machined) and after that KT value concedes the development of VBmax. Both these parameters VB and KT should be considered and taken into account in the model as main factors defining the remaining tool life.

Contact us now
 Home  Whatsapp  E-Mail  Inquiry